Wound infections remain a common and serious complication of anesthesia and surgery. Oxidative killing is the most important immune defense against surgical pathogens. Oxygen is a substrate both for this process and for proline hydroxylation of collagen (scar formation). Because wounds disrupt vascular supply and trigger clotting, they are hypoxic compared to normal tissue. Resistance to infection and wound repair can thus potentially be improved by increasing arterial oxygen tension, even of fully saturated blood. We propose a randomized trial evaluating the effects of supplemental oxygen administration on the incidence of surgical wound infections, duration of hospitalization, and cost of care. Patients undergoing elective colon resection under isoflurane anesthesia will be randomly assigned to 30 percent oxygen or 80 percent oxygen (balance nitrogen) intraoperatively and for two postoperative hours. The incidence of wound infection will be evaluated by a blinded observer, the decision to discharge patients from the hospital will be made by the (blinded) attending surgeon. The effects of supplemental oxygen on subcutaneous oxygen tension and collagen deposition will be evaluated in a subset of patients to help identify the factors impairing wound healing. We have so far evaluated 200 patients using this protocol. Oxygen saturation on the first postoperative day was virtually identical in the two groups, suggesting that supplemental oxygen did not cause clinically-important atelectasis. There were seven surgical wound infections in the patients given 80 percent oxygen and twelve in those given 30 percent (P = 0.38, Chi Square). The duration of hospitalization was longer in the patients given 30 percent oxygen (P = 0.15 by Wilcoxon test). Sample size analysis indicates that 996 patients will provide a 95 percent chance of identifying a statistically significant difference between the groups at a two-tailed alpha level of 0.05. We thus expect to demonstrate that providing supplemental perioperative oxygen (at a cost of five cents) roughly halves the incidence of surgical wound infections, impairs wound healing, and significantly reduces the duration and cost of hospitalization. Since the cost of oxygen is trivial and the risks of oxygen administration are low, clinicians will easily be able to implement this treatment once we have demonstrated its benefits.